wip: narrow deep paths now creates a chain of var edges (test included)

Author: nandikaK

gp_learner.py

@@ -413,19 +413,49 @@ def mutate_del_triple(child):
         return new_child
 
 
-def mutate_expand_node(child, pb_en_out_link):
-    # TODO: can maybe be improved by sparqling
-    nodes = list(child.nodes)
-    node = random.choice(nodes)
+def _mutate_expand_node_helper(node, pb_en_out_link=config.MUTPB_EN_OUT_LINK):
     var_edge = gen_random_var()
     var_node = gen_random_var()
     if random.random() < pb_en_out_link:
         new_triple = (node, var_edge, var_node)
     else:
         new_triple = (var_node, var_edge, node)
+    return new_triple, var_node
+
+
+def mutate_expand_node(child, node=None):
+    # TODO: can maybe be improved by sparqling
+    if not node:
+        nodes = list(child.nodes)
+        node = random.choice(nodes)
+    new_triple, _ = _mutate_expand_node_helper(node)
     return child + (new_triple,)
 
 
+def mutate_deep_narrow_path(
+        child,
+        min_len=config.MUTPB_DN_MIN_LEN,
+        max_len=config.MUTPB_DN_MAX_LEN,
+        term_pb=config.MUTPB_DN_TERM_PB,
+):
+    assert isinstance(child, GraphPattern)
+    nodes = list(child.nodes)
+    start_node = random.choice(nodes)
+   # target_nodes = set(nodes) - {start_node}
+    gp = child
+    hop = 0
+    while True:
+        if hop >= min_len and random.random() < term_pb:
+            break
+        if hop >= max_len:
+            break
+        hop += 1
+        new_triple, var_node = _mutate_expand_node_helper(start_node)
+        gp += [new_triple]
+        start_node = var_node
+    return gp
+
+
 def mutate_add_edge(child):
     # TODO: can maybe be improved by sparqling
     nodes = list(child.nodes)
@@ -647,7 +677,6 @@ def mutate(
         pb_ae=config.MUTPB_AE,
         pb_dt=config.MUTPB_DT,
         pb_en=config.MUTPB_EN,
-        pb_en_out_link=config.MUTPB_EN_OUT_LINK,
         pb_fv=config.MUTPB_FV,
         pb_id=config.MUTPB_ID,
         pb_iv=config.MUTPB_IV,
@@ -678,7 +707,7 @@ def mutate(
         child = mutate_del_triple(child)
 
     if random.random() < pb_en:
-        child = mutate_expand_node(child, pb_en_out_link)
+        child = mutate_expand_node(child)
     if random.random() < pb_ae:
         child = mutate_add_edge(child)
 
@@ -694,7 +723,6 @@ def mutate(
     else:
         children = [child]
 
-
     # TODO: deep & narrow paths mutation
 
     children = {

tests/test_gp_learner_offline.py

@@ -13,6 +13,7 @@
 from gp_learner import mutate_increase_dist
 from gp_learner import mutate_merge_var
 from gp_learner import mutate_simplify_pattern
+from gp_learner import mutate_deep_narrow_path
 from graph_pattern import GraphPattern
 from graph_pattern import SOURCE_VAR
 from graph_pattern import TARGET_VAR
@@ -108,6 +109,17 @@ def test_mutate_merge_var():
         assert False, "merge never reached one of the cases: %s" % cases
 
 
+def test_mutate_deep_narrow_path():
+    p = Variable('p')
+    gp = GraphPattern([
+        (SOURCE_VAR, p, TARGET_VAR)
+    ])
+    child = mutate_deep_narrow_path(gp)
+    assert gp == child or len(child) > len(gp)
+    print(gp)
+    print(child)
+
+
 def test_simplify_pattern():
     gp = GraphPattern([(SOURCE_VAR, wikilink, TARGET_VAR)])
     res = mutate_simplify_pattern(gp)
@@ -271,3 +283,7 @@ def test_remaining_gain_sample_gtps():
 
 def test_gtp_scores():
     assert gtp_scores - gtp_scores == 0
+
+
+if __name__ == '__main__':
+    test_mutate_deep_narrow_path()